The present invention relates to methods of operating droplet deposition apparatus, in particular an inkjet printhead, comprising a chamber communicating with a nozzle for ejection of ink droplets and with a supply of ink, the printhead further comprising electrically actuable means associated with the chamber and actuable a plurality of times to eject a corresponding number of droplets. In particular, it relates to a printhead in which the chamber is a channel having associated with it means for varying the volume of the channel in response to an electrical signal.
Such apparatus is known, for example, from WO95/25011, U.S. Pat. No. 5,227,813 and EP-A-O 422 870 (all incorporated herein by reference) and in which the channels are separated one from the next by side walls which extend in the lengthwise direction of the channels. In response to electrical signals, the channel walls are displaceable transverse to the channel axis. This in turn generates acoustic waves that travel along the channel axis, causing droplet ejection as is well-known in the art.
The last of the aforementioned documents discloses the concept of xe2x80x9cmultipulse greyscale printingxe2x80x9d: firing a variable number of ink droplets from a single channel within a short period of time, the resulting xe2x80x9cpacketxe2x80x9d of droplets merging in flight and/or on the paper to form a correspondingly variable-size printed dot on the paper. FIG. 1 is taken from the aforementioned EP-A-O 422 870 and illustrates diagrammatically droplet ejection from ten neighbouring printhead channels ejecting varying numbers (64,60,55,40,etc.) of droplets. The regular spacing of successive droplets ejected from any one channel indicates that the ejection velocity of successive droplets is constant. It will also be noted that this spacing Is the same for channels ejecting a high number of droplets as for channels ejecting a low number of droplets.
In the course of experiment, two deviations from the behaviour described in EP-A-O 422 870 have been discovered.
The first finding is that the first droplet to be ejected from a given channel is slowed by air resistance and may find itself hit from behind by subsequent droplets in the packet travelling In Its slipstream and therefore subject to less air drag. First and subsequent droplets of the packet may then merge to form a single, large drop.
The second finding is that the velocity of such a single, large drop will vary depending on the total number of droplets in the packet that are ejected in one go from a given channel.
A third finding relates to three-cycle operation of the printheadxe2x80x94described, for example in EP-A-O 376 532xe2x80x94in which successive channels in a printhead are alternately assigned to one of three groups. Each group is enabled in turn, with enabled channels ejecting a packet of one or more droplets in accordance with incoming print data as described above. It has been discovered that the velocity of the single, large drop formed by the merging of such droplets will vary depending on whether the adjacent channel in the same group is also being operated (i.e. 1 in 3 channels) or whether only the next-but-one channel in the same group is being operated (i.e. 1 in 6 channels).
The variations in velocity outlined above can give rise to significant dot placement errors which, although a known problem per se, can be particularly critical in printheads operating in the multipulse greyscale mode explained above. Here the present inventors have established that a placement error between two or more printed dots that is above one quarter of a pixel pitch can lead to print defects that are detectable by the naked eye. Since multipulse greyscale printheads typically operate at a printing pitch of 360 dots per inch and minimum substrate speeds, packet firing frequencies and printhead-substrate separations of 5 m/s, 5kHz and 1 mm respectively, this places an upper limit of 1.25 m/s on the acceptable variation in speed between the droplets that go to form any two adjacent printed dots.